Abstract
Background:
After the Centers for Medicare and Medicaid Services modified reimbursement rates for outpatient peripheral endovascular intervention (PVI) in 2008 with the intent of improving access to care, providers began to increasingly perform PVIs in privately-owned office-based clinics. Little is known about the characteristics of patients treated in this setting and their long-term outcomes as compared with those treated in hospital-based centers.
Methods:
In this retrospective cohort study, Medicare beneficiaries ≥66 years undergoing femoropopliteal PVIs in office-based clinics and hospital-based centers from 2015-2017 were identified. Sociodemographics, comorbidities, and institutional characteristics were compared across sites. Cox proportional hazard models were used to estimate the associations between variables and outcomes. The primary outcome was amputation-free survival (AFS) analyzed through the end of follow-up.
Results:
Among 134,869 patients, 29.9% were treated in office-based clinics and 70.1% in hospital-based centers. Patients treated in office-based clinics were more often Black (16.9% vs 11.9%), dually-enrolled in Medicaid (26.3% vs 19.6%) and residents of lower resourced regions (32.6% vs 25.6%). Over a median follow-up time of 800 days (IQR, 531-1119 days), patients treated in office-based clinics had improved adjusted AFS compared with those treated in hospital-based centers (HR 0.92, 95%CI: 0.89-0.95). They also had lower adjusted all-cause mortality (HR 0.93; 95%CI: 0.90-0.96), major lower extremity amputation (HR, 0.84; 95%CI: 0.79-0.89) and all-cause hospitalization (HR, 0.86; 95%CI: 0.84-0.88). These findings persisted after stratification by critical limb ischemia, race, dual enrollment and regional SES, as well as among operators treating patients in both clinical settings.
Conclusions:
In this large nationwide analysis of Medicare beneficiaries, office-based clinics treated a more socioeconomically disadvantaged population compared with hospital-based clinics. Long-term outcomes were comparable between locations. As such, these clinics appear to be selecting lower-risk patients for outpatient PVI, although there remains the possibility of unmeasured confounding.
Keywords: peripheral endovascular revascularization, peripheral artery disease, outcomes
Introduction
Peripheral arterial disease (PAD) affects nearly 12 million Americans with combined annual costs exceeding $21 billion1-3. As the standard in the treatment of symptomatic PAD has evolved from surgical bypass to endovascular revascularization, there has been growth in the establishment of privately-owned outpatient clinics that perform peripheral endovascular intervention (PVI)4. This shift was further motivated by the Centers for Medicare and Medicaid Services (CMS)’ decision in 2008 to modify payments for PVIs performed in the outpatient setting, either at hospital-based centers or privately-owned office-based clinics. The goal of this policy was to reduce costs, improve procedural efficiency and broaden access to care5. Consequently, 70% of PVIs shifted to outpatient centers6. However, these changes also resulted in greater variation in procedural practice patterns across settings. For instance, atherectomies, highly-reimbursed adjunctive devices that lacks high-quality randomized trial data, are disproportionately used in office-based clinics compared with other clinical settings7.
Older retrospective studies of the Medicare population have reported differences in 1-year outcomes among patients treated in office-based clinics compared with hospital-based settings7, 8. However, there has been significant growth in privately-owned office-based clinics in recent years, and little is known about the characteristics of the patients being treated at these sites. There is growing concern that office-based clinics may be disproportionately caring for racial/ethnic minorities and low-income adults and delivering lower quality care. Understanding whether the proliferation of office-based clinics has led to the delivery of inequitable care for vulnerable populations is critically important, and could inform policy strategies that aim to enhance equity in PAD care.
Therefore, this study evaluated a contemporary Medicare population undergoing PVI in order to: 1) characterize the demographics, race/ethnicity, and socioeconomic status of patients undergoing femoropopliteal PVI across different clinical settings; 2) determine long-term outcomes across clinical settings, accounting for differences in patient, procedural and institutional characteristics; and 3) understand how operators practicing in multiple clinical settings triage patients.
Methods
Data Availability
The data, analytic methods and study materials will not be made available to other researchers for purposes of reproducing the results or replicating the procedure.
Data Sources
Patients who underwent hospital-based or office-based PVIs from 4/1/2015 to 12/31/2017 were identified using 100% samples of the Institutional Outpatient files and cohort-specific Carrier fee-for-service files. The study was conducted in accordance with a prespecified statistical analysis plan and in compliance with the data use agreement that is in place between the CMS and Beth Israel Deaconess Medical Center. The study was approved by the institutional review board of Beth Israel Deaconess Medical Center, with a waiver of informed consent for retrospective data analysis.
Study Population
All Medicare fee-for-service beneficiaries ≥66 years who underwent femoropopliteal PVI identified by Current Procedural Terminology (CPT) and International Classification of Diseases (ICD)-10 claims codes were included in the study (Figure S1). Inclusion of these patients in the study required ≥1 year of Medicare claims prior to their index PVI to allow for ascertainment of comorbidities. We excluded patients with less than 1 year of Medicare claims. In addition, patients who did not have a claim in the Carrier file were excluded, as this was needed to determine place of service. For patients with multiple femoropopliteal PVIs within the study period, only the first procedure was selected. Patients were allowed to undergo simultaneous revascularization of lesions outside this target vessel during the same procedure.
Patient, Procedural and Institutional Characteristics
Race is self-reported to CMS and characterized as Black, White and Other. Dual enrollment in Medicare and Medicaid was used to identify beneficiaries experiencing poverty9. Linkage of the Medicare files with the Area Deprivation Index (ADI) was performed to identify socioeconomic status (SES) at the population level. The ADI is a composite measure of neighborhood socioeconomic disadvantage, with an ADI of <15th percentile representing the lowest resourced areas10-12.
Patient demographics and dual enrollment in Medicare and Medicaid were obtained from the Master Beneficiary Summary File. Patient comorbidities were ascertained using the chronic conditions included in the Master Beneficiary Summary File13 (Table S1). Additional claims codes were used to identify PAD severity (i.e. claudication and critical limb ischemia [CLI]) and prior amputation14 (Table S2). The type of index PVI procedure (e.g. stent +/− angioplasty, angioplasty alone, stent + atherectomy, or angioplasty + atherectomy) was identified using procedure claims codes (Table S3)14. Center characteristics included geographical region.
Primary Exposure
The primary exposure was the clinical setting of the index PVI. Place of service codes were used from Carrier file claims to differentiate hospital-based centers (codes 19 and 22) versus office-based clinics (i.e. ambulatory surgical centers and office-based labs; codes 11 and 24). Hospital-based centers included procedures performed in either an off-campus hospital based department that provides interventional services or within a hospital but does not require hospitalization. To confirm the accurate categorization of office-based clinics, a random selection of 100 billing addresses identified as office-based clinics in the Carrier file were searched online by two investigators (AR and EAS). All addresses confirmed the accurate designation of these clinic types.
Outcomes
The primary outcome was amputation-free survival (AFS) after the index procedure. AFS was defined as time (in days) from the discharge date to the date of death or first major ipsilateral or contralateral amputation. Although patients with intermittent claudication infrequently experience amputation, this was chosen as the primary endpoint as it is meaningful for all patients with peripheral artery disease undergoing revascularization. It is also a standard outcome reported in randomized-controlled trials and retrospective cohort studies and highlighted as a potential risk for patients treated with specific interventional devices14-18. Secondary outcomes after the index procedure included repeat PVI, surgical revascularization, major lower extremity amputation, minor lower extremity amputation and all-cause hospitalization (Table S4). Repeat PVI and surgical revascularization involved femoropopliteal artery procedures to either leg, as outpatient coding does not allow for identification of laterality.
Statistical Analysis
Categorical variables were reported as counts and percentages, and continuous variables as means with standard deviations. To test for differences between groups, χ2 tests for categorical variables and t tests for continuous variables were used.
The Kaplan-Meier method was used to calculate the cumulative incidence of AFS, and the log-rank test to evaluate for differences between groups. For all other outcomes, the cumulative incidence of events was calculated based on estimates of the cumulative incidence function, which accounts for the competing risk of mortality, and differences between groups evaluated using Gray tests.
Multivariable Cox regression models were used to calculate hazard ratios (HR) comparing outcomes by clinical settings. In all Cox regressions, to account for the intraclass correlations among patients from the same hospital, marginal models with a robust sandwich estimate of the covariance matrix were calculated, and statistical inferences were based on a robust sandwich estimate of the variances. Models were adjusted for age, sex, race, comorbidities, PAD indication, dual Medicare-Medicaid enrollment, ADI, procedural factors and institutional characteristics. Variables most strongly associated with AFS based on their risk estimates in the models were identified across all settings; a hazard ratio greater than 1 represented a higher risk of major amputation or death. Fine-Gray methods were used for all non-death outcomes to account for competing risks. As an alternative statistical approach to account for confounding, we also conducted the analysis using inverse probability of treatment weighting.
Subgroup analyses stratified by clinical setting were performed among patients of Black race, residing in the lowest ADI regions (ADI <15th percentile) and with dual enrollment. For each comparison, adjustment was performed using the same variables as in the primary analyses; however, for each comparison, risk estimates were calculated both with and without adjustment for other measures of SES/race.
A separate subgroup analysis was performed among patients with a diagnosis of CLI, as CLI is more prevalent among minority and lower SES patients. Similar statistical methods were used as in the primary analysis. Additionally, to better understand whether socio-demographics were associated with triaging patients to a particular outpatient setting, an analysis of patient characteristics and outcomes among providers practicing in both office-based clinics and hospital-based centers was performed.
Statistical significance was established by reporting 95% confidence intervals and using an α < 0.05. All analyses were performed using SAS 9.4 (Cary, NC, USA).
Results
Patient and Institutional Characteristics
Among 134,869 patients who underwent femoropopliteal PVI, 40,345 (29.9%) were treated in office-based clinics and 94,524 (70.1%) in hospital-based centers. For the total population, the mean (SD) age was 77.3 (7.4) years and 44.7% were female. Patients treated in office-based clinics had a higher rate of diabetes, hyperlipidemia, hypertension and stroke/transient ischemic attack than those in hospital-based centers (Table 1). They were also more likely to be Black (16.9% vs 11.9% hospital-based), dually-enrolled in Medicaid (26.3% vs 19.6% hospital-based) and residents of lower resourced regions (32.6% vs 25.6% hospital-based).
Table 1.
Baseline characteristics of patients admitted for index procedure, by clinical settings
| Number (%) | ||||
|---|---|---|---|---|
| Characteristic | Total (n = 134869) |
Office- Based Procedures (n = 40345) |
Hospital- Based Procedures (n = 94524) |
P value |
| Age, mean (SD) | 77.2 (7.3) | 77.4 (7.3) | 76.9 (7.3) | <.001 |
| Female | 60292 (44.7) | 18279 (45.3) | 42013 (44.4) | .004 |
| Race/ethnicity | <.001 | |||
| White | 108267 (80.3) | 29776 (73.8) | 78491 (83.0) | |
| Black | 18075 (13.4) | 6812 (16.9) | 11263 (11.9) | |
| Other | 8527 (6.3) | 3757 (9.3) | 4770 (5.0) | |
| Comorbidities | ||||
| Acquired hypothyroidism | 21443 (15.9) | 6798 (16.8) | 14645 (15.5) | <.001 |
| Acute myocardial infarction | 3052 (2.3) | 862 (2.1) | 2190 (2.3) | .04 |
| Alzheimer’s disease | 4165 (3.1) | 1540 (3.8) | 2625 (2.8) | <.001 |
| Alzheimer’s disease and related disorders or senile dementia | 14078 (10.4) | 4989 (12.4) | 9089 (9.6) | <.001 |
| Anemia | 48829 (36.2) | 16025 (39.7) | 32804 (34.7) | <.001 |
| Asthma | 7142 (5.3) | 2324 (5.8) | 4818 (5.1) | <.001 |
| Atrial fibrillation | 18366 (13.6) | 5199 (12.9) | 13167 (13.9) | <.001 |
| Benign prostatic hyperplasia | 13865 (10.3) | 4715 (11.7) | 9150 (9.7) | <.001 |
| Breast cancer | 3042 (2.3) | 948 (2.3) | 2094 (2.2) | .13 |
| Cataract | 25418 (18.8) | 7683 (19.0) | 17735 (18.8) | .23 |
| Chronic kidney disease | 49946 (37.0) | 15984 (39.6) | 33962 (35.9) | <.001 |
| Chronic obstructive pulmonary disease and bronchiectasis | 31301 (23.2) | 9173 (22.7) | 22128 (23.4) | .007 |
| Claudication | 45879 (34.0) | 15976 (39.6) | 29903 (31.6) | <.001 |
| Colorectal cancer | 2057 (1.5) | 572 (1.4) | 1485 (1.6) | .04 |
| Congestive heart failure | 39590 (29.4) | 12071 (29.9) | 27519 (29.1) | .003 |
| Critical limb ischemia | 34117 (25.3) | 9659 (23.9) | 24458 (25.9) | <0.01 |
| Depressive disorders | 21308 (15.8) | 6549 (16.2) | 14759 (15.6) | .004 |
| Diabetes | 70938 (52.6) | 22285 (55.2) | 48653 (51.5) | <.001 |
| Endometrial cancer | 310 (0.2) | 81 (0.2) | 229 (0.2) | 0.15 |
| Glaucoma | 14645 (10.9) | 4825 (12.0) | 9820 (10.4) | <.001 |
| Hip/pelvic fracture | 1218 (0.9) | 363 (0.9) | 855 (0.9) | 0.93 |
| Hyperlipidemia | 89863 (66.6) | 27644 (68.5) | 62219 (65.8) | <.001 |
| Hypertension | 108188 (80.2) | 33121 (82.1) | 75067 (79.4) | <.001 |
| Ischemic heart disease | 83033 (61.6) | 24761 (61.4) | 58272 (61.6) | .34 |
| Lung cancer | 2098 (1.6) | 591 (1.5) | 1507 (1.6) | .09 |
| Osteoporosis | 8693 (6.4) | 2719 (6.7) | 5974 (6.3) | .004 |
| Other peripheral artery disease | 37007 (27.4) | 7625 (18.9) | 29382 (31.1) | <.001 |
| Prior amputation | 2127 (1.6) | 337 (0.8) | 1790 (1.9) | <.001 |
| Prostate cancer | 5924 (4.4) | 1881 (4.7) | 4043 (4.3) | .002 |
| Rheumatoid arthritis/osteoarthritis | 52163 (38.7) | 16506 (40.9) | 35657 (37.7) | <.001 |
| Stroke/transient ischemic attack | 9722 (7.2) | 3216 (8.0) | 6506 (6.9) | <.001 |
| Measures of socioeconomic status | ||||
| Dual Medicaid-Medicare enrollment | 29080 (21.6) | 10594 (26.3) | 18486 (19.6) | <.001 |
| Area Deprivation Index | <.001 | |||
| <15th percentile | 37351 (27.7) | 13164 (32.6) | 24187 (25.6) | |
| ≥15th percentile | 97518 (72.3) | 27181 (67.4) | 70337 (74.4) | |
| U.S. geographic region | <.001 | |||
| Northeast | 4941 (3.7) | 791 (2.0) | 4150 (4.4) | |
| Middle Atlantic | 15742 (11.7) | 4461 (11.1) | 11281 (11.9) | |
| Northeast Central | 20527 (15.2) | 3100 (7.7) | 17427 (18.4) | |
| Southeast Central | 10441 (7.7) | 2417 (6.0) | 8024 (8.5) | |
| Northwest Central | 9004 (6.7) | 1340 (3.3) | 7664 (8.1) | |
| Southwest Central | 18518 (13.7) | 5889 (14.6) | 12629 (13.4) | |
| South Atlantic | 33011 (24.5) | 14401 (35.7) | 18610 (19.7) | |
| Mountain | 5980 (4.4) | 1288 (3.2) | 4692 (5.0) | |
| Pacific | 16396 (12.2) | 6529 (16.2) | 9867 (10.4) | |
| Other | 309 (0.2) | 129 (0.3) | 180 (0.2) | |
| Type of index intervention | <.001 | |||
| Stent +/− angioplasty | 30597 (22.7) | 3234 (8.0) | 27363 (28.9) | |
| Angioplasty alone | 35695 (26.5) | 4200 (10.4) | 31495 (33.3) | |
| Stent + atherectomy | 23273 (17.3) | 13322 (33.0) | 9951 (10.5) | |
| Angioplasty + atherectomy | 44851 (33.3) | 19642 (48.7) | 25209 (26.7) | |
| Simultaneous revascularization of iliac artery | 17271 (12.8) | 5146 (12.8) | 12125 (12.8) | .72 |
| Simultaneous revascularization of tibial artery | 37283 (27.6) | 15647 (38.8) | 21636 (22.9) | <.001 |
CI indicates confidence interval; HR, hazard ratio; and PVI, peripheral vascular intervention.
Institutional and procedural characteristics also varied across settings. Office-based clinics were more often located in the South Atlantic, Pacific and Southwest Central regions. Procedures at office-based clinics more often involved adjunctive atherectomy (81.7% vs 37.2% hospital-based) (Table 1).
Outcomes Following Treatment at Office-Based Clinics Versus Hospital-Based Centers
The median follow-up after PVI was 800 days (IQR, 531-1119 days). When comparing office-based clinics with hospital-based centers, the unadjusted cumulative incidence of major lower extremity amputation or death was lower (4-year frequency of 33.0% vs 36.2%). The unadjusted incidences of major lower extremity amputation and all-cause hospitalization were also lower (Figure 1). The unadjusted incidence of repeat PVI was higher in office-based clinics. Following adjustment, patients treated in office-based clinics continued to have higher AFS (HR 0.92, 95%CI: 0.89-0.95; P<.001) and lower hazards of all-cause mortality (HR 0.93; 95%CI: 0.90-0.96; P <.001), major lower extremity amputation (HR, 0.84; 95%CI: 0.79-0.89; P <.001), minor lower extremity amputation (HR, 0.78; 95%CI: 0.73-0.83; P<.001) and all-cause hospitalization (HR, 0.86; 95%CI: 0.84-0.88; P <.001) (Table 2). Patients treated in office-based clinics continued to have a higher likelihood of repeat PVI after adjustment (HR 1.52; 95%CI: 1.46-1.59; P <.001). Similar findings were ascertained using the inverse probability of treatment weighting methodology (Table S5).
Figure 1. Kaplan-Meier plots of the incidence of unadjusted outcomes, by clinical settings.
Kaplan-Meier curve showing higher unadjusted amputation-free survival over time among patients treated in office-based clinics compared with those treated in hospital-based centers (A). Kaplan-Meier curve showing a lower unadjusted risk of major lower extremity amputation over time among patients treated in office-based clinics compared with those treated in hospital-based centers (B). Kaplan-Meier curve showing a higher unadjusted risk of repeat endovascular revascularization over time among patients treated in office-based clinics compared with those treated in hospital-based centers (C). Kaplan-Meier curve showing a lower unadjusted risk of all-cause hospitalization over time in patients treated in office-based clinics compared with those treated in hospital-based centers (D).
CI indicates confidence interval; HR, hazard ratio; and PVI, peripheral vascular intervention.
Table 2.
Adjusted models for all outcomes, by clinical settings
| Outcomes | Office-Based Clinics vs. Hospital-Based Centers | |
|---|---|---|
| Adjusted HR (95% CI)*† | P value | |
| Major lower extremity amputation or death | 0.92 (0.89-0.95) | <.001 |
| All-cause mortality | 0.93 (0.90-0.96) | <.001 |
| Major lower extremity amputation | 0.84 (0.79-0.89) | <.001 |
| Minor lower extremity amputation | 0.78 (0.73-0.83) | <.001 |
| Repeat PVI | 1.52 (1.46-1.59) | <.001 |
| Surgical revascularization | 1.02 (0.93-1.13) | 0.634 |
| All-cause hospitalization | 0.86 (0.84-0.88) | <.001 |
CI indicates confidence interval; HR, hazard ratio; and PVI, peripheral vascular intervention.
Model adjusted for age, sex, race, comorbidities, PAD indication, procedural factors, institutional characteristics, dual Medicare-Medicaid enrollment, and ADI.
Generalized estimating equation method with robust variance estimator was performed to account for intraclass correlation of patients from the same hospital.
Among a subgroup of patients with a diagnosis of CLI, the findings were overall unchanged, with comparable adjusted outcomes following treatment at office-based clinics compared to hospital-based centers (Table S6).
Relationship between Race and SES with AFS across Clinical Settings
Across both clinical settings, dual-enrollment status and Black race were independently associated with decreased AFS (HR 1.17; 95%CI: 1.14-1.19; P <.001; HR 1.04; 95%CI: 1.02-1.07, P<.001, respectively) (Table S7).
Among Black patients, treatment at office-based clinics versus hospital-based settings was associated with higher adjusted AFS and lower risks of mortality and amputation. These findings persisted irrespective of additional adjustment for ADI and dual enrollment (Figure 2). Similarly, among patients residing in the lowest SES regions, treatment at office-based clinics versus hospital-based settings was associated with comparable outcomes after adjustment. This relationship persisted following additional adjustment for race and dual enrollment. Lastly, among dually-enrolled patients, treatment at office-based clinics versus other settings was associated with comparable outcomes. Risk estimates did not significantly differ with and without adjustment for race and ADI.
Figure 2. Forest plots of adjusted outcomes among total cohort and cohorts stratified by ADI, dual-enrollment status and Black race.
Among the total cohort and cohorts stratified by ADI, dual-enrollment status and Black race, patients treated in office-based clinics had higher amputation-free survival (A) and lower hazards of all-cause mortality (B) and major lower extremity amputation (C).
*Total: Model adjusted for age, sex, race, comorbidities, PAD indication, procedural factors, institutional characteristics, dual Medicare-Medicaid enrollment and ADI.
†ADI: Model adjusted for age, sex, race, comorbidities, PAD indication, procedural factors, institutional characteristics and dual Medicare-Medicaid enrollment.
‡Dual status: Model adjusted for age, sex, race, comorbidities, PAD indication, procedural factors, institutional characteristics and ADI.
§Black race: Model adjusted for age, sex, comorbidities, PAD indication, procedural factors, institutional characteristics and dual Medicare-Medicaid enrollment.
ADI indicates Area Deprivation Index; CI, confidence interval; and PAD, peripheral artery disease.
Providers Practicing in Both Office-Based Clinics and Hospital-Based Centers
Among 837 providers who performed PVIs in office-based clinics, 329 (39.3%) also submitted claims for procedures performed in hospital-based centers. When comparing characteristics of patients treated by these providers, those who underwent PVIs in office-based clinics were more often Black, dually-enrolled in Medicaid and residing in lower resourced regions compared with patients treated at hospital-based centers (Table S8). The unadjusted cumulative incidence of AFS was higher while incidences of all-cause mortality and major amputation were lower among patients treated in office-based clinics compared with those in hospital-based centers, whereas repeat PVI was performed more often among those treated in office-based clinics (Figure S2, Table S9). After adjustment, these differences in outcomes remained significant (Table S10).
Discussion
In this contemporary analysis of more than 130,000 Medicare beneficiaries who underwent femoropopliteal PVI between 2015 and 2017, we found significant variation in patient characteristics and long-term outcomes across clinical settings. Patients treated in office-based clinics had a greater burden of comorbidities compared with hospital-based centers, including being more often Black, dually enrolled in Medicaid and residents of socioeconomically disadvantaged regions. Nonetheless, these patients treated in office-based clinics had comparable outcomes compared with those treated in hospital-based centers, suggesting the potential selection of overall lower-risk patients for office-based treatment. These findings persisted among subgroups of patients of Black race and lower SES, as well as those with CLI. Repeat PVI did occur more often among patients treated in office-based clinics and of minority status, and represents the possibility of procedural overuse.
CMS’ changes in the outpatient prospective payment system were intended to both improve care access and reduce costs. However, by incentivizing the establishment of privately-owned centers, they also created financial benefits that could potentially influence procedural practices to optimize personal gain. For instance, drug-coated balloons, which have been recommended by societies as first-line therapies during femoropopliteal PVI19, are rarely used in office-based clinics due to the additional costs of the device and lack of incremental reimbursement compared with uncoated balloons20. Conversely, atherectomy, which is reimbursed at a level that exceeds the cost of the device7, is excessively used in office-based clinics.
With this in mind, we hypothesized that minority race and decreased SES would influence treatment location and portend worse long-term outcomes. Disparities have long plagued patients with PAD, with disparate rates of amputations among minority patients21 and reduced rates of revascularization attempts22, 23. Similar to other conditions and procedures, disadvantaged patients may be more likely to be targeted for alternative service sites24.
As suspected, we found substantial heterogeneity in baseline characteristics of patients treated across clinical settings. This was highlighted by more patients of Black race, higher comorbidity burdens and lower SES being revascularized in office-based clinics. These findings persisted among providers who practiced in both office-based clinics and hospital-based centers. Although the burden of comorbidities was greater among patients treated in office-based clinics, this may not represent a patient’s full procedural risk. For instance, more claudicants were treated in privately-owned clinics, and this condition carries lower procedural and long-term risks.
In this analysis, we found that long-term outcomes after PVI were comparable among patients treated at office-based clinics. Similarly, among patients of Black race, dually enrolled in Medicaid and residents of lower SES regions, outcomes were comparable at office-based clinics compared with hospital-based centers. This differs from a previous study, which preceded a major shift of PVIs into office-based clinics7. Notably, this finding was reproduced among all patients, stratified by treatment site, and similarly among operators treating patients in both clinical settings. Our interpretation of the data is that the comparable outcomes in these centers were likely driven by the selection of lower-risk patients for outpatient treatment, consistent with the expectations of these clinics. Although we do not conclude these office-based clinics provide better care due to the high likelihood of unmeasured confounding, these results may suggest that providers practicing in these centers are identifying lower-risk patients for treatment.
One concerning finding in this analysis was the greater use of repeat PVI among patients treated in office-based clinics. In part, this may be due to the risk profiles of these patients and the decreased need for surgical revascularization and amputation that are only performed in inpatient centers. However, this should be monitored closely. As these centers continue to grow in number, there needs to be vigilance to ensure that equitable outcomes and care delivery persist. The wide variability in patient characteristics also underscores the need to provide ongoing evaluation of patient selection and outcomes to ensure all patients are provided high quality care, irrespective of clinical site. This may be driven by a system that aligns payment more closely with quality and outcomes. Another ongoing effort is mandating the reporting of procedural data to a national quality registry, which can assist in determining institutional outliers and standardizing outcomes25.
Limitations
The limitations of the study include its observational design and the possibility of residual bias. In particular, unmeasured factors that can determine the treatment site selection and outcomes may not be completely captured in a claims-based analysis, such as lesion complexity, frailty of the patient and distance to the treatment site. In addition, this analysis lacks granularity with respect to baseline procedural characteristics, such as information about device specifications. Moreover, although the finding of greater use of repeat PVI among patients treated in office-based clinics warrants further investigation, we are limited by our inability to distinguish whether a repeat PVI or amputation involved the target vessel/lesion or the contralateral limb. The dataset also lacks information about patient-reported outcomes, hemodynamic changes pre- and post-PVI and medication use at baseline and follow-up periods. In addition, ~40% of patients lacked data to define the severity of their PAD, as outpatient procedures do not include exhaustive comorbidity claims codes and PAD is not ascertained easily in administrative billing data. There is also the potential for misclassification error with the use of claims codes, although this influence may be attenuated by the fact that these claims codes are linked to compensation and have been validated14. Additionally, Medicare fee-for-service beneficiaries represent an older population of patients with higher rates of comorbidities, which may reduce the generalizability of results. However, Medicare insures a significant proportion of patients with PAD given its increasing prevalence with age. Finally, the dataset cannot be used to identify cause-specific mortality.
Conclusions
In this large nationwide analysis of Medicare beneficiaries, we found that patients undergoing femoropopliteal revascularization procedures at office-based clinics were more socioeconomically disadvantaged, of minority race and had a higher burden of comorbidities compared with hospital-based centers. Nevertheless, long-term risks were comparable between settings. Overall, office-based clinics appear to be treating more low-risk, minority and low SES patients. Further efforts are needed to better monitor access to care and quality to ensure equitable outcomes for all patients with PAD.
Supplementary Material
What is Known
As the standard in the treatment of peripheral artery disease has shifted to endovascular revascularization, there has been a rise in the number of privately-owned outpatient clinics that perform these interventions.
This shift was also influenced by the Centers for Medicare and Medicaid Services’ 2008 decision to modify payments for outpatient-based interventions.
What the Study Adds
In this nationwide analysis of Medicare beneficiaries, patients undergoing femoropopliteal revascularization procedures at office-based clinics were more socioeconomically disadvantaged, of minority race and had a higher burden of comorbidities compared with those treated in hospital-based outpatient centers.
Long-term outcomes including major amputation-free survival or death, major amputation and all-cause hospitalization were comparable between settings, although rates of repeat revascularization were higher in office-based clinics.
These findings suggest that further efforts are needed to better monitor access to care and quality to ensure equitable outcomes within this patient population.
Sources of Funding
Dr. Secemsky is funded in part by NIH/NHLBI 1K23HL150290-01A1.
Disclosures
Relationships with Industry
RKW: Abbott
JFF: Research Grants: Commonwealth Fund, Robert Wood Johnson Foundation, Arnold Ventures, Harvard Center for AIDS Research.
RWY: Consulting: Abbott Vascular, AstraZeneca, Boston Scientific, Medtronic, Shockwave Medical, Zoll. Research Grants: AstraZeneca, Abbott Vascular, Bard, Boston Scientific, Cook, Medtronic, Philips.
EAS: Research grants: NIH/NHLBI K23HL150290, Harvard Medical School’s Shore Faculty Development Award, AstraZeneca, BD, Boston Scientific, Cook, CSI, Laminate Medical, Medtronic and Philips. Consulting/Speaking: Abbott, Bayer, BD, Boston Scientific, Cook, CSI, Inari, Janssen, Medtronic, Philips, and VentureMed.
All other authors have nothing to disclose.
Abbreviations
- ADI
Area Deprivation Index
- AFS
Amputation-free survival
- CI
Confidence interval
- CMS
Centers for Medicare and Medicaid Services
- CPT
Current Procedural Terminology
- ICD
International Classification of Diseases
- PAD
Peripheral artery disease
- PVI
Peripheral vascular intervention
- SES
Socioeconomic status
References
- 1.Belch JJ, Topol EJ, Agnelli G, Bertrand M, Califf RM, Clement DL, Creager MA, Easton JD, Gavin JR 3rd, Greenland P, et al. Critical issues in peripheral arterial disease detection and management: A call to action. Arch Intern Med. 2003;163:884–892 [DOI] [PubMed] [Google Scholar]
- 2.Muir RL. Peripheral arterial disease: Pathophysiology, risk factors, diagnosis, treatment, and prevention. J Vasc Nurs. 2009;27:26–30 [DOI] [PubMed] [Google Scholar]
- 3.Mahoney EM, Wang K, Cohen DJ, Hirsch AT, Alberts MJ, Eagle K, Mosse F, Jackson JD, Steg PG, Bhatt DL. One-year costs in patients with a history of or at risk for atherothrombosis in the united states. Circ Cardiovasc Qual Outcomes. 2008;1:38–45 [DOI] [PubMed] [Google Scholar]
- 4.Criqui MH, Alberts MJ, Fowkes FG, Hirsch AT, O'Gara PT, Olin JW, Writing Group 2. Atherosclerotic peripheral vascular disease symposium ii: Screening for atherosclerotic vascular diseases: Should nationwide programs be instituted? Circulation. 2008;118:2830–2836 [DOI] [PubMed] [Google Scholar]
- 5.Centers for medicare medicaid services. Medicare program; prospective payment system for long-term care hospitals ry 2007: Annual payment rate updates, policy changes, and clarification. Final rule. Fed Regist. 2006;71:27797–27939 [PubMed] [Google Scholar]
- 6.Creswell J, Abelson R. Medicare payments surge for stents to unblock blood vessels in limbs. New York Times. January 29, 2015. https://www.nytimes.com/2015/01/30/business/medicare-payments-surge-for-stents-to-unblock-blood-vessels-in-limbs.html [Google Scholar]
- 7.Jones WS, Mi X, Qualls LG, Vemulapalli S, Peterson ED, Patel MR, Curtis LH. Trends in settings for peripheral vascular intervention and the effect of changes in the outpatient prospective payment system. J Am Coll Cardiol. 2015;65:920–927 [DOI] [PubMed] [Google Scholar]
- 8.Turley RS, Mi X, Qualls LG, Vemulapalli S, Peterson ED, Patel MR, Curtis LH, Jones WS. The effect of clinical care location on clinical outcomes after peripheral vascular intervention in medicare beneficiaries. JACC Cardiovasc Interv. 2017;10:1161–1171 [DOI] [PubMed] [Google Scholar]
- 9.Wadhera RK, Wang Y, Figueroa JF, Dominici F, Yeh RW, Joynt Maddox KE. Mortality and hospitalizations for dually enrolled and nondually enrolled medicare beneficiaries aged 65 years or older, 2004 to 2017. JAMA. 2020;323:961–969 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hu J, Kind AJH, Nerenz D. Area deprivation index predicts readmission risk at an urban teaching hospital. Am J Med Qual. 2018;33:493–501 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kind AJH, Buckingham WR. Making neighborhood-disadvantage metrics accessible - the neighborhood atlas. N Engl J Med. 2018;378:2456–2458 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Singh GK. Area deprivation and widening inequalities in us mortality, 1969-1998. Am J Public Health. 2003;93:1137–1143 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chronic Conditions Data Warehouse. Centers for Medicare & Medicaid Services. Accessed September 2, 2022. https://www2.ccwdata.org/web/guest [Google Scholar]
- 14.Secemsky EA, Raja A, Shen C, Valsdottir LR, Schermerhorn M, Yeh RW, SAFE-PAD investigators. Rationale and design of the safe-pad study. Circ Cardiovasc Qual Outcomes. 2021;14:e007040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Adam DJ, Beard JD, Cleveland T, Bell J, Bradbury AW, Forbes JF, Fowkes FGR, Gillepsie I, Ruckley CV, Raab G, et al. Bypass versus angioplasty in severe ischaemia of the leg (basil): Multicentre, randomised controlled trial. Lancet. 2005;366:1925–1934 [DOI] [PubMed] [Google Scholar]
- 16.Conte MS, Bandyk DF, Clowes AW, Moneta GL, Seely L, Lorenz TJ, Namini H, Hamdan AD, Roddy SP, Belkin M, et al. Results of prevent iii: A multicenter, randomized trial of edifoligide for the prevention of vein graft failure in lower extremity bypass surgery. J Vasc Surg. 2006;43:742–751; discussion 751 [DOI] [PubMed] [Google Scholar]
- 17.Katsanos K, Spiliopoulos S, Paraskevopoulos I, Diamantopoulos A, Karnabatidis D. Systematic review and meta-analysis of randomized controlled trials of paclitaxel-coated balloon angioplasty in the femoropopliteal arteries: Role of paclitaxel dose and bioavailability. J Endovasc Ther. 2016;23:356–370 [DOI] [PubMed] [Google Scholar]
- 18.Kumar A, Brooks SS, Cavanaugh K, Zuckerman B. Fda perspective on objective performance goals and clinical trial design for evaluating catheter-based treatment of critical limb ischemia. J Vasc Surg. 2009;50:1474–1476 [DOI] [PubMed] [Google Scholar]
- 19.Klein AJ, Jaff MR, Gray BH, Aronow HD, Bersin RM, Diaz-Sandoval LJ, Dieter RS, Drachman DE, Feldman DN, Gigliotti OS, et al. SCAI appropriate use criteria for peripheral arterial interventions: An update. Catheter Cardiovasc Interv. 2017;90:E90–E110 [DOI] [PubMed] [Google Scholar]
- 20.Shishehbor MH, Jaff MR, Beckman JA, Misra S, Schneider PA, Lookstein R, Kashyap VS, Aronow HD, Jones WS, White CJ. Public health impact of the centers for medicare and medicaid services decision on pass-through add-on payments for drug-coated balloons: A call to action. JACC Cardiovasc Interv. 2018;11:496–499 [DOI] [PubMed] [Google Scholar]
- 21.Mustapha JA, Fisher BT Sr, Rizzo JA, Chen J, Martinsen BJ, Kotlarz H, Ryan M, Gunnarsson C. Explaining racial disparities in amputation rates for the treatment of peripheral artery disease (pad) using decomposition methods. J Racial Ethn Health Disparities. 2017;4:784–795 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Holman KH, Henke PK, Dimick JB, Birkmeyer JD. Racial disparities in the use of revascularization before leg amputation in medicare patients. J Vasc Surg. 2011;54:420–426 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Hughes K, Boyd C, Oyetunji T, Tran D, Chang D, Rose D, Siram S, Cornwell E 3rd, Obisesan T. Racial/ethnic disparities in revascularization for limb salvage: An analysis of the national surgical quality improvement program database. Vasc Endovascular Surg. 2014;48:402–405 [DOI] [PubMed] [Google Scholar]
- 24.U.S. Census Bureau, Statistical Abstract of the United States: 2012. Table 168. Ambulatory care visits to physicians’ offices and hospital outpatient and emergency departments: 2008. U.S. National Center for Health Statistics, National Health Statistics Reports. 2012. http://www.cdc.gov/nchs/ahcd.htm. [Google Scholar]
- 25.Ahn SS, Tahara RW, Jones LE, Carr JG, Blebea J. Preliminary results of the outpatient endovascular and interventional society national registry. J Endovasc Ther. 2020;27:956–963 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data, analytic methods and study materials will not be made available to other researchers for purposes of reproducing the results or replicating the procedure.